Multivariate Fuzzy Logic Based Smart Healthcare Monitoring for Risk Evaluation of Cardiac Patients

Abstract

In this chapter, the multi-dimensional fuzzy logic design is implemented to develop smart and pervasive healthcare application systems for determining the risk status of cardiac patients. This heuristic decision-making and fuzzy reasoning mechanism is based on the medical observations of several miscellaneous physiological conditions of systolic blood pressure, blood sugar level, serum cholesterol, resting heart rate and body mass index. These disparate attributes controlling the heart risk factor are jointly contributed as the input variables to the presented fuzzy inference unit. The deployed sample record comprising of the particular heart patients' health data informatics is effectively enforced by the fuzzy control scheme and is consequently analyzed through the rigorous simulation experiments. Wide range of error and fairness indicators are employed for corroborating the accuracy estimation of the fuzzy output parameter of heart disease risk level. Specifically, we applied the normalized root-mean-square error, mean Poisson deviance, absolute logarithmic quotient error, relative absolute error, symmetric mean absolute deviation, entropy and Jain's fairness index for empirical substantiation of the proposed method in relation to the actual heart disease observational statistics. From the simulation analysis results, it was established that the optimal solution is obtained with the normalized root-mean-square error of 0.459 and mean Poisson deviance of 0.244 error approximations. In addition, the absolute log quotient error, relative absolute error, and symmetric mean absolute deviation measures are accordingly computed as 0.73, 4.6, and 0.65. Besides, the Shannon entropy and fairness indices for the statistical analysis of the proposed soft computing based biomedical application are relatively improved by almost 3.337 and 13.24% as compared to the actual heart disease dataset under surveillance.